Model-Agnostic Counterfactual Reasoning for Eliminating Popularity Bias in Recommender System

Abstract

The general aim of the recommender system is to provide personalized suggestions to users, which is opposed to suggesting popular items. However, the normal training paradigm, i.e., fitting a recommender model to recover the user behavior data with pointwise or pairwise loss, makes the model biased towards popular items. This results in the terrible Matthew effect, making popular items be more frequently recommended and become even more popular. Existing work addresses this issue with Inverse Propensity Weighting (IPW), which decreases the impact of popular items on the training and increases the impact of long-tail items. Although theoretically sound, IPW methods are highly sensitive to the weighting strategy, which is notoriously difficult to tune. In this work, we explore the popularity bias issue from a novel and fundamental perspective -- cause-effect. We identify that popularity bias lies in the direct effect from the item node to the ranking score, such that an item's intrinsic property is the cause of mistakenly assigning it a higher ranking score. To eliminate popularity bias, it is essential to answer the counterfactual question that what the ranking score would be if the model only uses item property. To this end, we formulate a causal graph to describe the important cause-effect relations in the recommendation process. During training, we perform multi-task learning to achieve the contribution of each cause; during testing, we perform counterfactual inference to remove the effect of item popularity. Remarkably, our solution amends the learning process of recommendation which is agnostic to a wide range of models -- it can be easily implemented in existing methods. We demonstrate it on Matrix Factorization (MF) and LightGCN [20]. Experiments on five real-world datasets demonstrate the effectiveness of our method.

To Appear in SIGKDD 2021

References in corpus (5)

• Causal Intervention for Leveraging Popularity Bias in Recommendation
• Clicks can be Cheating: Counterfactual Recommendation for Mitigating Clickbait Issue
• Degenerate Feedback Loops in Recommender Systems
• Should Graph Convolution Trust Neighbors? A Simple Causal Inference Method
• DeVLBert: Learning Deconfounded Visio-Linguistic Representations

Cited by in corpus (32)

• A Survey on Popularity Bias in Recommender Systems
• Generalizing to the Future: Mitigating Entity Bias in Fake News Detection
• Investigating Accuracy-Novelty Performance for Graph-based Collaborative Filtering
• A Stakeholder-Centered View on Fairness in Music Recommender Systems
• Comprehensive Fair Meta-learned Recommender System
• Unbiased Knowledge Distillation for Recommendation
• CausalRec: Causal Inference for Visual Debiasing in Visually-Aware Recommendation
• Towards Individual and Multistakeholder Fairness in Tourism Recommender Systems
• Causal Collaborative Filtering
• Debiasing Recommendation with Personal Popularity
• CausalMed: Causality-Based Personalized Medication Recommendation Centered on Patient health state
• Uncovering User Interest from Biased and Noised Watch Time in Video Recommendation
• How Do Recommendation Models Amplify Popularity Bias? An Analysis from the Spectral Perspective
• Dynamic Causal Collaborative Filtering
• Leveraging Watch-time Feedback for Short-Video Recommendations: A Causal Labeling Framework
• The Bandwagon Effect: Not Just Another Bias
• Balancing User Preferences by Social Networks: A Condition-Guided Social Recommendation Model for Mitigating Popularity Bias
• LabelCraft: Empowering Short Video Recommendations with Automated Label Crafting
• Cross-modal Variational Auto-encoder for Content-based Micro-video Background Music Recommendation
• Capturing Popularity Trends: A Simplistic Non-Personalized Approach for Enhanced Item Recommendation
• Test Time Embedding Normalization for Popularity Bias Mitigation
• Robust Basket Recommendation via Noise-tolerated Graph Contrastive Learning
• Automatic Feature Fairness in Recommendation via Adversaries
• Popularity Debiasing from Exposure to Interaction in Collaborative Filtering
• Connecting Domains and Contrasting Samples: A Ladder for Domain Generalization
• Causality-aware Graph Aggregation Weight Estimator for Popularity Debiasing in Top-K Recommendation
• Performative Debias with Fair-exposure Optimization Driven by Strategic Agents in Recommender Systems
• Countering Mainstream Bias via End-to-End Adaptive Local Learning
• Generating Negative Samples for Multi-Modal Recommendation
• SPARK: Adaptive Low-Rank Knowledge Graph Modeling in Hybrid Geometric Spaces for Recommendation
• Not All Videos Become Outdated: Short-Video Recommendation by Learning to Deconfound Release Interval Bias
• On Inherited Popularity Bias in Cold-Start Item Recommendation